Hack your Research! @ ECOC 2026
By Marco
Event Date: Tuesday, 22 September 2026, 5:45pm – 7:30pm CEST
Event Location: E1 Exhibition room 1, FYCMA, Málaga, Spain
Link: ecoc2026.org/ECOC2026/programme/special-events
Call for Demos:
Submit a demo proposal.
Organizers:
Besma Kalla, Eindhoven University of Technology, Netherlands
Giammarco Di Sciullo, Università degli Studi dell’Aquila, Italy
Amol Delmade, Pilot Photonics, Ireland
Advisory Committee:
Rekha Yadav, University College London, United Kingdom
Vincent van Vliet, Eindhoven University of Technology, Netherlands
Marco Eppenberger, PsiQuantum, USA
Demos:
Designing an experimental setup using CAD
by Omar Abdel-Aal, Eindhoven University of Technology, the Netherlands
This demo shows how to use Autodesk Inventor to create a 3D blueprint of your setup. Check tolerances, align optics, ensure components fit, and design custom 3D-printed parts which help you avoid headaches during assembly even before purchasing any components.
Making Wavelength Selective Switching Devices
by Lauren Dallachiesa, Nokia Bell Labs, USA
Come and look inside a wavelength selective switch and a channel blocker. We’ll talk through the shared architecture, the one design choice that separates them, the budgets that limit both, and the alignment, gluing and calibration that make a built device work.
From the Theory to the Lab: Utilizing Simulations to Guide Experimental Work in SDM Fibers
by Lucas Alves Zischler, University of L’Aquila, Italy
Many physical effects occur simultaneously in space-division-multiplexed fiber systems, making measurements challenging to interpret. Theory helps disentangle these effects and connect experimental observations to their physical origins. Using examples from non-linear measurements in multimode and multicore fibers, this demo illustrates how simulations and analytical models can aid studies in these novel fiber types.
Recirculating loops: go across the ocean comfortably sitting in your lab
by Stefano Gaiani, National Institute of Information and Communications Technology, Japan
The total available length of optical fibers in laboratories is always limited, which makes research on transmission over long-haul and transoceanic distances more complex. Here, we detail how to set up a multiband recirculating loop to perform experiments with single-mode and SDM fibers over hundreds or thousands of kilometers starting from just a few kilometers of fiber.
Lab Equipment Control with python
by Julian Schneck, University of Stuttgart, Germany
Annoyed by ever-changing communication protocols and funky quirks of your lab equipment? Feel overwhelmed by 42 different implementations from former students? Then check out this object-based implementation in Python. Based on ABC metaclasses, it forces every new implementation into a well-defined, common form.
Generate Vector Graphics with TikZ
by Felipe Villenas Cortez, Eindhoven University of Technology, the Netherlands
With just LaTeX and a text file with data, you can use TikZ to generate vector graphics such as plots and block diagrams. In this demo, I will show some of the main advantages of using TikZ for the figures in your papers, reports, and presentations, alongside some basic examples to get started.
GPU programming in Python and MATLAB for dummies
by Kadir Gümüş, Eindhoven University of Technology, the Netherlands
In this demo, I will give a quick overview of the basics of GPU programming for both Python and MATLAB. These basics are enough to significantly speed up any of your code, and even with little to no experience, you’ll be able to program on the GPU after this demo.
Learn how to configure and operate a MATLAB-based coherent transport DSP with up to 700 parameters using LabVIEW!
by Robert Emmerich, Fraunhofer Heinrich-Hertz-Institut, Germany
The recent advent of open-source software has led to a decline in the use of proprietary environments such as the graphical programming languages LabVIEW and MATLAB. However, these tools are still in use in state-of-the-art optical transmission labs and require continuous optimisation of operations. In this demonstration, we will present an interactive demo based on an experiment conducted in our Berlin laboratory. This demo will use our MATLAB-based coherent transport DSP in conjunction with a LabVIEW-based GUI for configuration and DSP control. Furthermore, an introduction to the setup and construction of the multi-band transmission system used for this year’s OFC post-deadline paper will be given to viewers.
The lost art of using GPUs not for AI
by Mindaugas Jarmolovičius, University College London (UCL), London, United Kingdom
Do you have a piece of code that just takes forever to run? No? Well, do you have a thesis to write and would just do anything else, let’s say, making your code run faster? If that’s the case, rewriting everything in C++ is (probably) not the procrastination you want. This demo will enlighten you with advanced procrastination methods making not only your number-crunching Python code run faster than C++ could, but also will shock you with how little GPU VRAM you can get way with while doing it. This demo will introduce basic and more advanced concepts of running Python code on GPUs (big and small, new and old ones), providing a guide through a jungle of GPU frameworks, mostly focusing on JAX ecosystem, while also encompassing discussions of why you may or may not want to do this in your research, and other Hack Your Research style tricks.
see https://github.com/zceemja/labathon_ecoc2026
Plug, Play, and Talk: Agentic Optical Networks with ETSI TeraFlowSDN
by Lluis Gifre, CTTC, Spain and Andrea Sgambelluri, CNIT/SSSA, Italy
This demonstration presents the latest enhancements in the ETSI TeraFlowSDN controller to manage optical transport networks through agentic automation. The demo combines recent optical-controller advances for ROADMs, Transponders, optical connectivity services, and flex-grid spectrum handling with new Model Context Protocol (MCP) and Agent-to-Agent (A2A) capabilities. TeraFlowSDN exposes its control functions through MCP, enabling AI agents to inspect topology, devices, links, services, and optical spectrum resources. Multiple TeraFlowSDN controller instances, each responsible for a different optical domain, coordinate through Agent-to-Agent (A2A) interfaces to negotiate cross-domain optical connectivity services. The demo leverages emulated optical devices, allowing experimentation without requiring specialized hardware. TeraFlowSDN is an open-source, cloud-native, and microservice-based SDN controller developed mainly in Python and hosted under the umbrella of ETSI. New contributors are welcome to join. This demo has been developed in the context of PROTEUS-6G (GA 101139134).
Splice: streaming telemetry and agentic control for mixed-vendor networks
by Agastya Raj, Tian Tian and Marco Ruffini, Trinity College Dublin, Ireland
Disaggregated, multi-vendor networks promise freedom from lock-in. But in practice every box speaks its own management dialect, so monitoring, control, and automation get rebuilt per vendor. Splice is a lightweight, open-source (Apache 2.0, pure Python) bridge that hides those differences behind one clean interface: a single client can configure a device, stream live telemetry, and be driven by an AI agent - whatever protocol (NETCONF, gNMI) the hardware underneath speaks. It’s the missing glue for a genuinely vendor-neutral stack, from a lab testbed up to a production disaggregated line system. At the table I’ll show it live: an agent driving mixed gear with telemetry streaming back in real time - grab the repo and point it at your own network.